NORTH ATLANTIC OSCILLATION CONTROL OF DROUGHTS IN NORTH-EAST SPAIN: EVALUATION SINCE 1600 A.D.
SERGIO M. VICENTE-SERRANO1 and JOSÉ M. CUADRAT2
1Instituto Pirenaico de Ecología, CSIC (Spanish Research Council), Campus de Aula Dei, P.O. Box 202, Zaragoza 50080, Spain
e-mail:
2 Departamento de Geografía. Universidad de Zaragoza. Zaragoza. Spain.
Abstract. This paper analyses the role played by the North Atlantic Oscillation (NAO) in the creation of drought conditions in a semi-arid region of North-east Spain (the middle Ebro valley), from 1600 to the year 2000. The study used documents from ecclesiastical archives for the seventeenth, eighteenth and nineteenth centuries. For the twentieth century, instrumental precipitation records were used as well. A December-August drought index from 1600 to 1900 was compiled from the historical documentary sources (rogation ceremonies). The index was validated by means of precipitation records between 1858 and 1900 and independent precipitation data from 1600 reconstructed by means of dendrochronological records. Using instrumental data a drought index was also calculated (Standardized Precipitation Index, SPI) for the 1958 – 2000 period. We found that the NAO was important in explaining the droughts identified in the study area from documents and instrumental data. Positive values of the winter NAO index are prone to cause droughts in the middle Ebro valley. This finding has been verified since 1600 by means of two independent reconstructions of the winter NAO index. The same behaviour has been observed during the nineteenth and twentieth centuries by means of instrumental records. The climatic and geographic factors that explain the high influence of North Atlantic Oscillation on droughts in this region are discussed in depth.
1. Introduction
Drought is a complex phenomenon which involves both human and natural factors. Although there are different drought types: agricultural, hydrological and socioeconomic (Wilhite and Glantz, 1985), it is usually understood to be a long and sustained period in which water becomes scarce and the scarcity causes negative impacts on the society and/or environment (Dracup et al., 1980; Redmond, 2002). Thus, drought can be considered to be essentially a climatic phenomenon related to an abnormal decrease in precipitation (Oladipo, 1985; McKee et al., 1993).
Drought is one of the main climatic hazards affecting Mediterranean regions. In Spain, it is a frequent phenomenon due to the high spatial and temporal variability of precipitation. Several studies have identified dry periods that have affected Spain in different centuries (Martín-Vide and Barriendos, 1995; Rodrigo et al., 1999; Saz, 2003). Droughts were also frequently recorded in Spain during the period when instruments were used (Pérez-Cueva, 1983; Pita, 1989; Vicente-Serrano, 2006).
The main characteristics of drought in Spain are: the spatial differences and the non-synchrony of droughts between different regions (Martín-Vide, 2001; Vicente-Serrano, 2006b), and the severity and duration of the episodes (Pérez-Cueva and Escrivá, 1982). The recent drought period between 1991 and 1995, which was the most intense of the twentieth century in Spain, is an excellent example of these characteristics (Almarza et al., 1999).
Droughts in Spain are determined by atmospheric circulation variability in the North Atlantic region (Rodriguez-Puebla et al., 1998; Trigo and Palutikof, 2001; Trigo et al., 2004; Olcina, 2001; Zorita et al., 1992; Xoplaki et al., 2004; Barriendos and Llasat, 2003; Martin-Vide and Fernández, 2001). Atmospheric circulation in the Northern Hemisphere can be characterised in terms of its patterns. Barnston and Livezey (1987) identified nine patterns in the Northern hemisphere during the boreal winter and three patterns in the other seasons. One pattern, called the North Atlantic Oscillation (NAO), occurs in all seasons though it is more intense in winter (Rogers, 1984; Hurrell, 1995; Trigo and Palutikof, 2001).
The NAO explains much of the climatic variability in the North Atlantic areas, including the direction and intensity of the westerlies, the trajectories of polar depressions and the position of anticyclones (Lamb and Peppler, 1987; Hurrell et al., 2003). Changes of atmospheric circulation in the North Atlantic are related to changes in the amount, distribution and intensity of precipitation events (Hurrell, 1995; Hurrell et al., 2003). Positive NAO winters usually show dry conditions in the Western Mediterranean areas (Hurrell and Van Loon, 1997; García Herrera et al., 2001; Xoplaki et al., 2004; Moses et al., 1987).
In Spain, it is well known that the NAO affects precipitation, but with important spatial differences (Trigo et al., 2004). The effect of the NAO on precipitation is stronger in winter but it is also detectable in other seasons such as spring and autumn (Martín-Vide and Fernández, 2001). The NAO mainly affects precipitation in the South-Atlantic basins (Rodríguez-Puebla et al., 1998; Martín-Vide and Fernández, 2001). In other areas of Spain precipitation is determined by other atmospheric patterns, such as the ENSO (Rodó et al., 1997; Vicente-Serrano, 2005), the Scandinavian (Rodríguez-Puebla et al., 1998) or the Polar pattern (González-Hidalgo et al., 2003; Xoplaki et al., 2004). Nevertheless, the NAO not only affects precipitation over large areas of central and South-western Spain; in the North-east as well there is a small region in which precipitation is highly affected by it (Martín-Vide and Fernández, 2001; Vicente-Serrano and López-Moreno, 2006). This area has semi-arid climatic characteristics (Cuadrat, 1999), in which the temporal variability of droughts is very important (Vicente-Serrano and Cuadrat, 2006). Moreover, droughts are very severe (it is common to have more than 100 consecutive days with zero precipitation, Vicente-Serrano and Beguería, 2003). Droughts have an important effect on the economy and environment of the region because drought variability determines crop production (Austin et al., 1998; Vicente-Serrano et al., 2006) and the development of natural vegetation (Vicente-Serrano, 2004; Creus and Saz, 2004).
The objectives of this paper are to find out how drought evolves in this semi-arid region, and to analyse in depth the impact of the NAO on droughts. The purpose is to determine whether the NAO influence on droughts can be recognised over wide periods. The NAO’s effect on droughts from 1600 to 2000 was analysed by means of instrumental records and proxy data from historical documentary sources.
The paper is structured as follows: Chapter 2 explains the characteristics of precipitation in the middle Ebro valley. Chapter 3 describes the data and the methods used. In chapter 4 the main results of this research are discussed. This section has been divided into two sub-sections: the first explains the temporal variability of droughts in the middle Ebro valley since the Seventeenth century and the second shows the relationships between the NAO and drought variability. Finally, section 5 provides the main conclusions of the paper.
2. The climate of the region
The principal city of the Ebro valley is Zaragoza (600,000 inhabitants) (Figure 1). This city is in the centre of one of the most arid regions of Europe. Annual precipitation is 343.3 mm, although the interannual variability is very high (standard deviation = 86 mm). Using instrumental records, the minimum annual precipitation since 1858 was recorded in 1912 (182.6 mm) whilst the maximum was in 1959 (752 mm). The highest evapotranspiration rates are recorded in the centre of the valley (> 1200 mm/year). In the whole of the region there is a negative water balance (precipitation – evapotranspiration), which is particularly significant in the central areas (> 900 mm).
The climate is Mediterranean with important continental characteristics (Köppen, 1936). It is determined by various geographic factors, as several mountainous chains, which isolate the valley from moist winds, border the Ebro valley. So in the central areas of the valley precipitation is low (Cuadrat, 1999; Creus and Ferraz, 1995), with little differences between months, although the dry seasons are the summer and the winter (mean winter precipitation between 1901 and 2000 = 63.8 mm (standard deviation = 34.3), mean spring precipitation = 96.6 mm (49.0), mean summer precipitation = 65.4 mm (38.4) and mean autumn precipitation = 93.6 mm (49.4)).
3. Data bases and methods
3.1. CLIMATIC RECORDS FROM HISTORICAL DOCUMENTARY SOURCES
Ecclesiastical documents from the sixteenth century, which have a high degree of detail and no gaps in information, are conserved in Zaragoza (see appendix). The conservation of the documents during this long period (four centuries) is due to the duplication of the ecclesiastical archive because of the existence of two ecclesiastical sites: “La Seo” and “El Pilar”. In Zaragoza there are two cathedrals as a consequence of the confrontation between the chapters of the basilicas of El Pilar and la Seo de El Salvador. The dispute was resolved in the 17th century by a papal decision, which granted the title of cathedral to both basilicas, and the two chapters were joined together, but for several years two ecclesiastical archives were maintained, which were also finally combined in 1776. Each site retained one general register, in which each document of the other site was duplicated to avoid internal conflicts in the ecclesiastical community. This has conserved the general archive without gaps from 1557 to the present.
Numerous documents giving information about meteorological characteristics appear in these historical documentary sources. The records refer to rogation ceremonies taking place due to environmental factors. The rogations “pro pluvia” and “pro serenitate” have frequently been used in Catholic countries as “proxy” climatic data (Martin-Vide and Barriendos, 1995; Barriendos, 1996-1997; Barriendos, 1997; Barriendos and Martín-Vide, 1998; Piervitali and Colacino, 2001). Figure 2 gives an example from the ecclesiastical documentary sources conserved in Zaragoza.
Rogations are a part of Catholic liturgy. They are solemn petitions by believers to ask God specific requests (Barriendos, 1996-1997). The “pro-pluvia” rogations were made to ask for precipitation during a drought and, therefore, they provide an indication of episodes of drought. On the other hand, the “pro-serenitate” rogations asked for the end of the precipitation during periods of high and/or intense precipitation, which caused floods and damages. In the Mediterranean region the most frequent economic problem, the loss of crops, was related to insufficient rainfall. For this reason, in a period of important religious fervour, the rogations “pro pluvia” were the most frequent human response to climatic anomalies, and fortunately these are well recorded in documentary sources. In this paper we focussed on the “pro-pluvia” rogation records, which were a social response to the drought events.
The credibility of the rogations, as a manifestation of a meteorological event, is supported by the participation of several institutions (agricultural organisations, municipal and ecclesiastical authorities) that analysed the situation and deliberated before deciding to hold a rogation ceremony. Agricultural organisations requested the rogations when a drop in rainfall impeded crop development. Municipal authorities recognised the problem and discussed the advisability of holding a rogation ceremony. The order was communicated to the religious authorities, who placed the rogation on the calendar of religious celebrations and finally organised and announced the rogation.
The rogation ceremonies might vary between localities (Barriendos, 1996-1997), but in all cases had the advantage of the formality characteristic of the Catholic Church’s ceremonies. In Zaragoza five rogation levels, depending on the seriousness of the meteorological event, are recorded (Table 1). When the level increases, more complex ecclesiastical ceremonies were performed and greater drought severity can be inferred. In Zaragoza, the first level of rogations corresponded to a single petition for precipitation during the masses of different churches. In the second level there was a direct petition for precipitation to a saint of the church. If the precipitation did not occur the rogations gained in complexity. The third level was characterised by different masses and processions within the church and a public procession in the neighbouring streets with a holy image or the relics of a saint. The fourth level consisted of litanies, masses and a public procession that usually included the churches of El Portillo, Santa Engracia and the Hospital de Nuestra Señora de Gracia. The most frequent relic used in the procession was the head of San Valero, patron saint of Zaragoza, which was sometimes put into the water to ask for precipitation. Finally, the fifth level consisted of masses in the cathedrals and a public procession with the Christ figure of La Seo de San Salvador, which was taken from this cathedral to the cathedral of El Pilar, where the image was left for some days in front of the image of the Virgin of El Pilar. At the same time, several litanies and masses were celebrated in other churches.
To carry out this study, we consulted the historical documentary sources in the cathedral archive of Zaragoza, which records rogation ceremonies from 1557 on. 98 books were reviewed. Finally, we decided to work with the records obtained from the beginning of the seventeenth century, as prior to 1588 rogation records are scarce.
The last rogation ceremony was recorded in 1945, but the reliability of twentieth-century records is low. Only 14 rogations between 1900 and 1945 were recorded, although important droughts were identified with instrumental data during this period (Vicente-Serrano, 2005b). The decline in the social and political influence of the Catholic Church during the first decades of the twentieth century, along with the instrumental records that are available, makes it inadvisable to use rogation records for drought quantification during this period. Rogations were used for drought analysis until 1900, but instrumental records from 1858 on were also used.
3.2. CALCULATION OF DROUGHT INDEX FROM ROGATION RECORDS
Quantitative continuous monthly series from 1600 to 1900 were created from the rogation records. We used the number of level 1, 2 and 3 rogations as a preliminary measurement of drought occurrence, and we selected the rogations from December to August for further analysis.
Different studies have shown that winter precipitation is the most important for the final crop productions in the dry-farming areas of the middle Ebro valley (wheat and barley) (Austin et al., 1998; McAneney and Arrúe, 1993) because water is stored in the soil as a consequence of the low evapotranspiration rates during the winter (McAneney and Arrúe, 1993; Austin et al., 1999). Therefore, it is reasonable to assume that the winter droughts would not just have consequences on the total number and the level of the rogations recorded in winter (December to March). The socio-economic effects would be more evident during the period of vegetation growth (March-May) and also during the harvesting period (June-July). For this reason, the rogations recorded between December and August were used to calculate an annual drought index and to relate it to the winter NAO.
We created a continuous drought index (DI) by grouping the rogations at various levels. We followed a simple approach, similar to that of Martín-Vide and Barriendos (1995). We did not use level 4 and 5 rogations because they are very scarce. They are concentrated in specific periods (consecutive years) and usually occur in years in which there were no lower-level rogations (79% of level 5 rogations were recorded in years with no level 1 rogations – 33% - or with only one level 1 rogation – 46%). This arouses suspicions as to the reliability of level 4 and 5 rogations for quantifying droughts, because they may have been held because of other social, economic or religious factors.
Annual values of the drought index (DI) were obtained by means of the weighted average of the number of level 1, 2 and 3 rogations recorded in the period between December and August. The weight for each level was 1, 2 and 3, respectively.
3.3. VALIDITY OF ROGATION RECORDS FOR DROUGHT QUANTIFICATION
Numerous authors attest to the validity of rogations in quantifying droughts (e.g. Pfister, 1999; Barriendos, 1997; Piervitali and Colacino, 2001; Barriendos, 2005; Luterbacher et al., 2006 and references therein). In general, the information is highly credible due to the strict control of the process by the Catholic Church and the costs involved in holding the ceremonies, borne by the civil authorities (Martín-Vide and Barriendos, 1995).
In this paper we also quantitatively analysed the validity of the rogation records for drought quantification by comparing the rogation records with the precipitation series available in the Zaragoza observatory. Precipitation records start in 1858. Between this year and 1900, we have 39 years of complete records. The years 1863, 1864, 1886 and 1887 are missing.
Also we compared the DI with an independent reconstruction of past climate. Dendrochronological reconstructions of annual precipitation (October to September), which start in 1570 were used (Saz, 2003). The precipitation records were available in two observatories of the centre of the Ebro valley: Haro (190 km west of Zaragoza) and Pallaruelo de Monegros (85 km North-East). A series was generated from average of annual records in each observatory, weighted for the distance to Zaragoza.
Figure 3 shows the mean December-August precipitation between 1858 and 1900, years in which a varying number of rogations of each level were recorded. For level 1 rogations mean December-August precipitation was lower in the years with two rogations (175 mm) than in the years with only one rogation (193.6 mm). Moreover, mean December-August precipitation in years without level 1 rogation records was higher (231 mm). With level 2 rogations there are also differences in December-August precipitation values between the years without rogations of this level (221.2 mm) and years that did record a rogation of this level (190 mm). Differences are more important in the analysis of level 3 rogations. The December-August mean precipitation during the years in which one rogation of this level was recorded is only 169.7 mm.
In addition, DI values were compared with the December-August precipitation in Zaragoza (Figure 4). Results show that the years with high DI values, such as 1870, 1876, 1878 and 1896, coincide with a major decrease in precipitation, whereas in years with DI values of 0 and 1 annual precipitation was, in general, higher. This can be seen in the years 1858, 1862, 1871-1874 and 1883-1892. The correlation between the two time series is negative and statistically significant (R = -0.34, p < 0.05). Although the amount of shared variance between the DI and the precipitation is only a few percent, the characteristic of the rogation records does not allow a continuous quantitative comparison between the DI and the precipitation because the DI is only indicative of the drought events but it is not sensitive to high precipitation values. The DI is bounded in 0. This value would be recorded the years with normal precipitation but also in some years with high precipitation values, which would not be accompanied with a decrease of the DI.